This paper proposes a methodology for the design of hard real-time component-based applications that relies on a resource reservation paradigm. It enables configuring the applications to satisfy their timing requirements before knowing the workload of the platform in which they are deployed. The assembler carries out the design and schedulability analysis of an application based on a virtual execution platform, which is defined as a set of virtual schedulable resources, and a set of restrictions imposed on their parameters that must be satisfied so that the application is schedulable. Afterwards, when the application is deployed on a concrete real execution platform, the deployer negotiates, by means of a resource reservation service, the support for the virtual resources needed by the application. If they are compatible with the current workload of the platform, they are accepted, and the application is installed and executed with the guarantee that its timing requirements will be met.